Dynamic models for nonstationary signal segmentation

Penny, William D ORCID: https://orcid.org/0000-0001-9064-1191 and Roberts, Stephen J. (1999) Dynamic models for nonstationary signal segmentation. Computers and Biomedical Research, 32 (6). pp. 483-502. ISSN 0010-4809

Full text not available from this repository. (Request a copy)


This paper investigates Hidden Markov Models (HMMs) in which the observations are generated from an autoregressive (AR) model. The overall model performs nonstationary spectral analysis and automatically segments a time series into discrete dynamic regimes. Because learning in HMMs is sensitive to initial conditions, we initialize the HMM model with parameters derived from a cluster analysis of Kalman filter coefficients. An important aspect of the Kalman filter implementation is that the state noise is estimated on-line. This allows for an initial estimation of AR parameters for each of the different dynamic regimes. These estimates are then fine-tuned with the HMM model. The method is demonstrated on a number of synthetic problems and on electroencephalogram data.

Item Type: Article
Additional Information: Copyright 1999 Academic Press.
Uncontrolled Keywords: algorithms,electroencephalography,hand,humans,markov chains,statistical models,movement,computer-assisted signal processing,sleep
Faculty \ School: Faculty of Social Sciences > School of Psychology
UEA Research Groups: Faculty of Social Sciences > Research Centres > Centre for Behavioural and Experimental Social Sciences
Depositing User: Pure Connector
Date Deposited: 23 Aug 2017 05:04
Last Modified: 20 Apr 2023 00:33
URI: https://ueaeprints.uea.ac.uk/id/eprint/64644
DOI: 10.1006/cbmr.1999.1511

Actions (login required)

View Item View Item